The purpose of gait analysis is to quantify a person's walking pattern and the neuromuscular strategies used by the body to produce the observed movement. A computerized video camera system measures the movement pattern and special platforms built into the floor measure forces acting on the feet. From the measured movements and forces, the mechanical work done by the body during walking can be estimated. Current techniques used to approximate work oversimplify the model of the body and its movements and introduce errors into the estimates. The purpose of our research is to investigate more complex models of the body to learn how these models affect mechanical work estimates. The study of 10 normal subjects that examined the work done at the ankle joint during gait has been published. The study employed a more sophisticated model of the ankle joint and included additional terms, previously ignored, in the mechanical work calculations. Data from another study which examined the mechanical work done at the knee joint has been completed and presented. A third study involving another 15 subjects produced a dramatically improved model of the foot. A final study, currently underway, is investigating the effect of dramatically different walking speeds on the mechanics of walking. It is hypothesized that two modes of walking exist rather than a continuum of a single walking strategy. The results of these studies suggest that a complete and numerically consistent mechanical characterization of human movement is feasible. The application of these analytic techniques to the evaluation of human motion will have implications for rehabilitation and surgical treatment in patient populations.